Pharmacoepidemiological Approach to the Predisposing Factors for Highly Active Antiretroviral Therapy Failure in an Hiv-Positive Cohort from Cordoba City (Argentina) 1995–2005

Introduction

The illness caused by the human immunodeficiency virus (HIV) has a chronic progression, leading to a final stage constituted by acquired immunodeficiency syndrome (AIDS)-related complex and AIDS, after a long-standing phase of clinical latency. ¹ HIV replicates continuously, depending on the evolution stage, into CD4-positive lymphocytes, causing their depletion and predisposing the patient to opportunistic infections and tumours. ²

The introduction of highly active antiretroviral therapy (HAART) has helped to restore cellular immunity, avoid resistance strains, delay the clinical progression and diminish mortality.^3,4 Patient adherence to the pharmacological regimens plays a critical role in controlling HIV. ⁵ HAART should be initiated when a marker disease occurs, the viral load (VL) exceeds 100,000 viral copies/mL and/or the CD4 count (CD4C) is below 200 cells/mL. These are effective markers of disease progression, therapy response and prognosis.^6,7 Although therapy has reached medical goals, other problems have appeared, mainly adverse medicament reactions (AMRs). ⁸ The frequency of drug secondary effects exceeds AIDS diagnostic events. ⁹ Since the lipodystrophic syndrome has been studied widely, ⁸ it is necessary to evaluate other manifestations. In this regard, the aim of this work was to study the role of AMRs in the success of HAART in an Argentinian cohort under ambulatory management with government-supplied drugs.

Methods

Results

Fifteen active principles were combined into 63 different schemes, namely (indication times, n), 3TC (lamivudine)–-DDI (didanosine)–ABV (abacavir) (1), 3TC–EFV (efavirenz)–ABV (1), 3TC–NFV (nefavirenz)–EFV–ABV (1), 3TC–NVP (nevirapine)–ABV (20), 3TC–NVP–EFV (1), 3TC–RTV (ritonavir)–SQV (saquinavir) (2), AZT (zidovudine)–3TC (2), AZT–3TC–ABV (11), AZT–3TC–ATZV (atazanavir) (1), AZT–3TC–DDI–D4T (stavudine)–NVP (1), AZT–3TC–EFV (12), AZT–3TC–IDV (indinavir) (7), AZT–3TC–NFV (3), AZT–3TC–NVP (39), AZT–3TC–NVP–NFV (1), AZT–3TC–RTV (1), AZT–3TC–RTV–IDV (2), AZT–3TC–RTV–LPV (lopinavir) (4), AZT–3TC–RTV–SQV (27), AZT–D4T–3TC–DDI–EFV–IDV–ABV (1), AZT–DDC (zalcitabine) (4), AZT–DDC–RTV (1), AZT–DDI (5), AZT–DDI–EFV (1), AZT–DDI–IDV (2), AZT–DDI–NVP (2), AZT–DDI–RTV (2), AZT–DDI–RTV–SQV (1), AZT–TFV (tenofovir)–RTV–LPV (1), D4T–3TC–ABV (1), D4T–3TC–DDI–EFV–SQV–RTV–LPV (1), D4T–3TC–EFV–SQV–RTV–LPV (1), D4T–3TC–IDV (9), D4T–3TC–NFV (1), D4T–3TC–NVP (9), D4T–3TC–NVP–NFV (1), D4T–3TC–RTV–SQV (1), D4T–3TC–SQV (1), D4T–DDI–ABV (3), D4T–DDI–EFV (3), D4T–DDI–IDV (1), D4T–DDI–NFV (4), D4T–DDI–NVP (6), D4T–DDI–NVP–NFV (1), D4T–DDI–RTV (1), D4T–DDI–RTV–LPV (1), D4T–EFV–ABV (2), D4T–NFV–ABV (1), D4T–NVP–ABV (3), D4T–RTV–SQV (6), D4T–RTV–SQV–ABV (1), DDC–ABV–APV (1), DDC–RTV–LPV–ABV (2), DDI–ABV–RTV–LPV (4), DDI–NVP–ABV (4), DDI–NVP–NFV (1), DDI–RTV–ABV–IDV (2), DDI–EFV–RTV–TNF–TPV (1), NVP–RTV–LPV–ABV (1), NVP–RTV–SQV (1), RTV–IDV–ABV (1), RTV–LPV–SQV (1) and RTV–SQV (16). Schemes were changed at least once in the 57% of cases due to AMRs, which were found with varying frequency and intensity among patients. Antiretroviral pharmacotoxicity was the main cause of altered drug regimen followed by drug TF and by the aim of simplification, without the influence of age and sex. The medical indication of different regimens schemes was done according to international consensus, ¹⁴ using government-supplied drugs. Concerning AMRs, gastroenterological intolerance (dyspepsia, nausea, vomiting and diarrhoea) and soft tissue rheumatism (myositis and tenosinovitis) were common effects for different combinations, whereas toxidermia was seen only with NVP and 3TC. Overall, the most frequent collateral effects were TF (14% analytical and 10% clinical failure), dyslipaemia (20%), abnormal haemogram (16%), rheumatism (8%), digestive problems (7%) and abnormal hepatogram (6%). Dyslipaemic patients showed an increase in plasma level of both triglycerides (2–6 times) and cholesterol (2–3 times) with a low high-density lipoprotein level. Abnormal haemogram was mainly represented by macrocytic anaemia, whereas hepatogram acquired a colestatic pattern (2–4-fold increase of alkaline phosphatase and γ-glutamyl transpeptidase) and injury markers (1.5–3-fold increase of transaminases, mainly glutamylpyruvate transaminase). On the other hand, hyperamilasemia and hyperuricemia were eventual findings without clinical relevance. Other findings were less frequent (e.g. toxidermia, peripheral neuropathy). The immune reconstitution syndrome was represented by Kaposi's sarcoma (0.5%) and atypical mycobacteriosis (0.5%).

Twenty-nine schemes exhibited TF at least once. After accepting the model (deviance gL = 1.02) it was seen that the gender had no effect. The drug combinations differed broadly in regard to chance of failure (Figure 1). Compared with AZT–3TC–NVP (with a >75% success), the pre-HAART schemes failed more frequently (OR 8-fold, 95% CI 1.6–41.2). Other unsuccessful treatments were those that included IDV (30-fold, OR CI 5.6–157), whose failure rate was greater than that pre-HAART (6-fold, P < 0.04) and those with retrotranscriptase inhibitor (RTI) and ≥3 protease inhibitors (PI) (11-fold, OR CI 1.9–69.5) (P < 0.01). Nonetheless, the last ones were applied to viral-resistant documented patients. When EFV was used instead of NVP, TF OR increased 2.31-fold with a 0.6–8.3 OR CI. The rest of the drugs were similar to the first scheme. Drug toxicity (T) and drug change rate (C) were analysed in relation to drug indication quantity (I), constituting the TC/(I² – 1) index, applied to integrate T/I and C/I quotients. This index distributed asymmetrically (assymetry coefficient [AC] = 1.72) with a mean ± SE of 0.71 ± 0.13 (0–3.33) and a median of 0.49 (P [25] 0.23; P [50] 0.4; P [75] 1.19). It exhibited a direct relationship with TF outcomes (2 ± 0.5, CI = 0.96–3, R² = 65%, CC = 70%, P < 0.002). The number 1 set in the denominator annulled estimations when a drug was used only once. The analytical failure was preceded by inadequate adherence to therapy (>95%), whereas this antecedent was absent in successful treatments, exhibiting a strong association between AMRs and TF either in analytical or clinical cases (P < 0.005).

OPEN IN VIEWER

Figure 1

Antiretroviral schemes with therapeutic failure (TF) (1995–2005, Cordoba, Argentina). The TF is separated in analytical (dark grey) and clinical (light grey). White areas represent successes

Discussion

Pharmacological intervention in HIV infection pursues longterm VL suppression, immune improvement, a minimal morbidity/mortality and an optimal quality of life. The laboratory parameters CD4R and VL are effective markers for assessing infection evolution and pharmacological response. Nonetheless, this illness presents a complex challenge due to the increasing number of antiretroviral drugs and the emerging new data about them.^19,20 In this respect, the issues in need of attention are more effective teaching, practice guidelines, unnecessary, inappropriate and toxic drug treatments and new research on problems such as HIV, development of new drugs from natural products, new information technology and risk–benefit analysis of drug therapy. ²¹ Regarding this, the AMRs represent a major issue, because they are very frequent and detrimental, being the main problem in patients with normal CD4R. ⁹

AMRs were the principal reason for regimen change, followed by TF and medical simplification. Sex and age did not influence their incidence and their heterogeneous manifestations. In this regard, the occurrence of digestive symptoms was lower than the values found in the available data, indicating a subregister of those problems. The frequency found for other AMRs was in agreement with general data. ¹⁴ Another factor which should be considered in detail is the immune reconstitution syndrome. It cannot be definitively catalogued because it implies either marker disease relapse or treatment responsiveness. Consequently, it cannot be considered as TF.

Among the combinations studied, the AZT–3TC–NVP scheme was effective including in pregnant women. On the other hand, the IDV-containing combination therapy frequently failed. Consequently its use should be considered carefully, at least as the first choice treatment, despite the fact that this disagrees with previous studies. ²² Patients with resistant viral strains did not always respond positively when RTIs were used with three of more PIs. Concerning this and the critical role of drug costs and treatment efficacies for optimal HIV drug management in resource-limited settings, ²³ different parameters have been developed to evaluate drug choices where therapeutic resistance is seen. ²⁴

As TF is an appalling eventuality in HIV treatment, its determinants should be detected and prevented. Also, there is no gold standard for adherence, so clinically useful measures for antiretroviral medication adherence are encouraged and in current development in order to relate it to patient conditions.^25,26 Moreover, patients who experienced AMRs have reported either strong beliefs that HIV infection would cause them future health problems or distrust in the therapeutic benefits of treatment. Beliefs about future HIV-related health concerns were associated with suboptimal dose adherence. ²⁷ In this work, poor patient adherence mostly preceded TF, AMR being another predisposing factor. Furthermore, the significant association between low adherence, drug toxicity and TF must be taken into account in order to reach the desired medical goals in HIV treatment.